Dataclass field constraints

[binaryDiv]

Currently it is already possible to implement field constraints in dataclasses by defining a post validation method (either __post_init__() in the dataclass or post_validate() in the validator), which can check custom constraints like "if at least one of the fields A and B is set, both must be set" or "the sum of integer fields A and B must be positive".

Example:

@validataclass
class ExampleClass:
    enable_something: bool = BooleanValidator()
    some_value: Optional[int] = IntegerValidator(), Default(None)

    def __post_init__(self):
        if self.enable_something is True and self.some_value is None:
            raise DataclassPostValidationError(field_errors={
                'some_value': RequiredValueError(reason='Must be set if enable_something is True.'),
            })

It would be nice to have some sort of way to declaratively define such constraints without writing multiple lines of code for each constraint, especially a shorter way to raise the appropriate DataclassPostValidationError exceptions.

Not quite sure how, though. Maybe using a class member called _constraints_ or similar. But how generic should this be? Maybe it's enough to restrict it to conditional requirements (e.g. simple "if field A exists, field B must exist"). questioning